Rename "GC Pauses" to "GC Collection Time" in Pulsar JVM dashboards¬

- source metric is not about GC pauses.¬
- the source metric is "jvm_gc_collection_seconds_sum" which is not the same as "GC pause"¬
- The metric is defined in Prometheus Java client¬
  - "Time spent in a given JVM garbage collector in seconds."¬
  - source code is https://github.com/prometheus/client_java/blob/e68daf23336eb5de7856df406eb1d497f51ad3be/simpleclient_hotspot/src/main/java/io/prometheus/client/hotspot/GarbageCollectorExports.java#L53¬
    - calls GarbageCollectorMXBean.getCollectionTime()¬
      - Javadoc https://docs.oracle.com/en/java/javase/11/docs/api/java.management/java/lang/management/GarbageCollectorMXBean.html#getCollectionTime()¬
        "Returns the approximate accumulated collection elapsed time in milliseconds."¬
     - Stackoverflow Q&A: https://stackoverflow.com/a/44686539¬

dashboards.kubernetes/jvm.json

@@ -428,7 +428,7 @@
       "thresholds": [],
       "timeFrom": null,
       "timeShift": null,
-      "title": "GC Pauses",
+      "title": "GC Collection Time",
       "tooltip": {
         "shared": true,
         "sort": 0,

dashboards.template/jvm.json.j2

@@ -437,7 +437,7 @@ This will make graph ledgends display the pod name instead of IP and port inform
       "thresholds": [],
       "timeFrom": null,
       "timeShift": null,
-      "title": "GC Pauses",
+      "title": "GC Collection Time",
       "tooltip": {
         "shared": true,
         "sort": 0,

dashboards/jvm.json

@@ -428,7 +428,7 @@
       "thresholds": [],
       "timeFrom": null,
       "timeShift": null,
-      "title": "GC Pauses",
+      "title": "GC Collection Time",
       "tooltip": {
         "shared": true,
         "sort": 0,